How I look at a new technology project
The set of questions I run any new request through, from an AI assistant to industrial analytics.
When a new request lands - "we want an AI assistant", "we need a dashboard", "pull data off the shop floor" - I almost always start from the same set of questions. They are boring. They help.
Block 1. Why this is needed
- What problem are we actually solving?
- Who will be working with this every day, and how?
- What changes if the solution does not exist?
- What are we measuring as success?
If the answers here are vague, there is no point going further. Any technology placed on top of a fuzzy goal only hides the problem.
Block 2. The data
- Which data does this need?
- Where does it physically live?
- What state is it in?
- Who owns its correctness?
- Can we pull it on a schedule, or only as a one-off?
This block usually surfaces what the business already half-suspects: there is less data than people thought, or it is in worse shape than people thought. That is not a reason to stop. It is a reason not to start the project under illusions.
Block 3. Architecture and integrations
- Which existing systems will be touched?
- Which integrations are needed?
- Which of them are cheap, and which are painful?
- What happens to the solution if one of the systems changes?
I rarely propose the most "elegant" architecture. More often I propose the most durable one.
Block 4. Security and accountability
- Which data in this solution is sensitive?
- Who gets access?
- What needs to be logged?
- What happens in the event of an incident?
- Who answers for it once I'm out of the project?
This block is the easiest to skip and the most expensive to recover later.
Block 5. Deliverability
- Who will use this?
- Who will support it?
- What skills does the in-house team have?
- What needs to happen for the solution to keep running without me?
A good system is not the one people admire on launch day. It is the one that still works a year later, when the consultant has long left.
Block 6. Cost and time
- How much on day one?
- How much in operation?
- How much would it cost to stop, if we realise the direction is wrong?
Cost of ownership is almost always the most important question, and the one people remember last.
Why bother
These six blocks do not produce a flashy slide. They sharply reduce the chance of the owner saying, a year later: "Technically everything is right, but somehow it is not working."
In most of those stories, the "somehow" was visible in the first round of questions. Nobody just stopped long enough to ask them.